Compile-Time Type Vs Run-Time Type - [OOP & Java #3]

I have a Paper object that contains 2 equals method.

class Paper{
    private int height;
    private int width;

    Paper(int height, int width){
            this.height= height;
            this.width = width;
        }
    @Override
    public boolean equals(Object obj){
        System.out.println("A");
            return this == obj;
    }
      //Overload
    public boolean equals(Paper p){
        System.out.println("B");
        return this == p; 
    }
}

What will be printed out for the following 6 equals method calls?

//set up
Paper paper = new Paper(297, 140);
Object object = paper;
// 1 to 3
object.equals( object );
object.equals( (Paper) object );
object.equals( paper );
// 4 to 6 
paper.equals( object );
paper.equals( (Paper) object );
paper.equals( paper );

---

---

Answer as follows:

// 1 to 3
object.equals( object );         // A
object.equals( (Paper) object ); // A
object.equals( paper );          // A
// 4 to 6 
paper.equals( object );           // A
paper.equals( (Paper) object );   // B
paper.equals( paper );            // B

Let's get to the analysis. Feel free to correct me if I am wrong...

---

Basic knowledge

	Compile-time	Run-time
Type	Left hand side of the assignment, fixed upon declaration	Right hand side of the assignment, vary as program runs
Binding	Static	Dynamic
Control	Restricts methods that can be called based on compile-time type	Methods to invoke are determined during run-time
Check for	Method overloading, more specific method will be invoked	Method overriding, overridden method will never be invoked

There are a total of 3 different possible outputs :

• None (Java Object's default equals method is invoked)
• A (Paper's equals overriding method is invoked)
• B (Paper's equals overloading method is invoked)

Let's draw a proper mental model of what is going on.

Explanation for the set-up

//set up
Paper paper = new Paper(297, 140);
Object object = paper;

• A variable named paper is created and declared to be of compile-time type Paper. The variable is also assigned to a new Paper object created using the Paper class constructor. The run-time type of the variable is also Paper.
• A variable named object is created and declared to be of compile-time type Object. The variable is also assigned to the created Paper object. The run-time type of the variable object is Paper.

Explanation for the 1st to 3rd method calls

// 1 to 3
object.equals( object );         // A
object.equals( (Paper) object ); // A
object.equals( paper );          // A

During compile-time, we look at the compile-time type. So in the above three cases, we have a variable named object that has compile-time type of Object. This fact tells us that the variable can execute any method as long as the method is included in the Object class.

At the same time, we check if there is any method overloading, where two or more methods have the same name but different method signatures. If there is method overloading, we will look for the most specific method within that class to call upon. This is static binding.

During run-time, we know that certain objects might have different type as compared to compile-time, hence we check if any method has been overridden and change our decision made during compile-time, and run the overriding method instead. This is also known as dynamic binding.

From our mental model above, we can see that the Object class has only one equals method. So,

• Method call 1 is safe. During compile-time, we loosely understand method call 1 as (Object) object.equals( (Object) object ). Since we do have the equals method within Object class, we know at the very least, equals method can be called from within Object class. Come run-time, we realize that object has a run-time type of Paper. We now check for any override (not overload!), and turns out that there is an overriding equals method within the child Paper class and hence that method is invoked and printed "A".
• Method call 2 is safe. We can see that due to typecasting, a Paper object is being passed into the equals method. Although we do not have a particular equals method in Object class that takes in Paper, we can treat the parameter as Object. This is because Paper is a subclass of Object (by default, all objects in Java inherit from Object). Hence, following the same logical reasoning of method call 1, the equals method within Paper class is invoked and printed "A".
• Method call 3 is safe. Reasoning follows method call 2.

Additional note:

• The ability to invoke the overriding method in the subclass instead of the overridden method in the parent class supports polymorphism. This means given objects of the same parent, we can call a particular method and invoke different implementations of that method within each child class.

Explanation for the 4th to 6th method calls

// 4 to 6 
paper.equals( object );           // A
paper.equals( (Paper) object );   // B
paper.equals( paper );            // B

With the knowledge from the first three method calls, it is much easier to understand the rest.

• Method call 4 is safe. We do have an equals method within the Paper class and looking at the method signature, we know that we will call the one that specifically takes in Object, which outputs "A".
• Method call 5 is safe. (Paper)object signals that we want the object to be used as Paper when this method is run. During compile-time, We see that the parameter is typecast into Paper, so we look for a specific method call that match this. We have the second equals method within Paper that specify that it takes in Paper and outputs "B".
• Method call 6 is safe. Reasoning is similar to method call 5.

Additional note:

• The only time that the overloaded method equals(Paper) can be invoked is when the compile-time type of the variable is Paper.

---

Conclusion

Understand the different results across run-time, due to inheritance and polymorphism, might be slightly confusing. However, this knowledge will allow us to better appreciate Objec-Oriented design later on. One example is that polymorphism and dynamic binding will make extension of code easier (minimize modification of client code when adding new a subclass).